Name the client and their site, and the playbook comes back built on the claims they have already published: every scenario scored against a printed rubric, every holding statement written and marked for who has to approve it.
Give Juma the client and their URL. The Flow reads the brand's own public surface first, pulling the sustainability figures, sourcing and nutrition claims, labelling language and data commitments they have actually published, and records each one with the page it came from and the date it was checked. Those claims become the scenarios, so the register names the real exposure instead of listing the same generic incidents every brand gets.
What comes back is a branded PDF playbook and an editable CSV register: scenarios scored on a printed likelihood and impact rubric, a four-level escalation matrix with triggers and update cadence, a first-60-minutes sheet, holding statements written per channel and marked pre-approved or counsel required, and a do-not-say table with the reason behind every line. 400+ marketing teams use Juma, and this is the kind of document a team writes once and then cannot find at 2am.
Go deeper on the claims audit on its own terms. The Flow takes every public commitment the brand has made, the climate figures, the sourcing and ingredient language, the nutrition statements and the labelling, and tests each one for how well it is substantiated on the page where it appears. Each claim comes back with the wording as published, the source URL, the date checked, what a journalist or regulator would ask for next, and whether that evidence is publicly available. Claims that rest on an unpublished internal study are flagged separately from claims backed by a linked assessment, because those two fail in very different ways.
Expand the claims exposure audit into its own document. For every public claim, give me the wording as published, the source URL, the date checked, the evidence a journalist or regulator would ask for next, and whether that evidence is publicly available today. Flag any claim whose only support is an unpublished internal study.
A playbook nobody has rehearsed is a document, not a capability. This step turns the highest-scoring scenario in the register into a 90-minute tabletop exercise: an opening inject, three timed escalations that make the facts worse in realistic ways, the decision points the team has to clear at each stage, and a scoring sheet covering time to first response, fact discipline, approval friction and channel control. The exercise names which role owns each decision, so the gaps that show up are specific. Teams usually find the same two: nobody knows who can approve a statement out of hours, and the holding line is written for press while the pressure is on social.
Turn the highest-scoring scenario into a 90-minute tabletop exercise: an opening inject, three timed escalations that worsen the facts, the decision each one forces, and a scoring sheet for time to first response, fact discipline, approval friction and channel control. Say which role owns each decision.
No register covers everything, and the scenario that lands is often the one nobody wrote down. Describe what has happened and the Flow drafts the first public line against the same rules as the rest of the playbook: what is confirmed, what is being investigated, what affected people should do, and when the next update comes. It returns the full channel set, a press line, a post inside the X character limit with the count shown, a LinkedIn version, a support macro and an internal note, then marks the statement pre-approved or counsel required and runs it against the do-not-say table before handing it over.
Here is a situation the register does not cover: [describe what has happened and what is confirmed]. Draft the first public holding statement for it, give me the press, X, LinkedIn, customer-support and internal versions, mark it pre-approved or counsel required, and check it against the do-not-say table before you hand it over.
A single playbook breaks the moment an incident crosses a border, because the regulator, the notification clock and the language all change. This step builds a market pack for each country the client sells in: which authority has to be told and how fast, the statutory deadlines that start running the moment something is confirmed, the local spokesperson role and who approves on their behalf, the retailer or distributor contacts that need a separate note, and the holding statements translated with the legal phrasing checked rather than translated word for word. Each pack states what it could not confirm, so nobody treats a gap as a cleared route.
Build a market pack for each country the client sells in: the regulator to notify and how fast, the statutory deadlines that start on confirmation, the local spokesperson role and who approves for them, the retailer or distributor contacts needing a separate note, and the holding statements localised with the legal phrasing checked. State anything you could not confirm.
Run one project per client. The playbook gets sharper every time the project knows more about how that client actually operates, and the same context carries into every other Flow the team runs for them. Add these once and the next crisis document starts from what the client has already approved rather than from public pages alone.
Claims and Substantiation Pack
The evidence behind the client's public claims: lifecycle assessments, certifications, nutrition data and the approved wording for each one. With this in the project, the claims audit checks published language against the real evidence instead of inferring what might sit behind it.
Crisis Contact and Approval List
Who holds each role, who can approve a public statement, who covers out of hours, and how each person is reached. The escalation matrix is built around these names and the approval paths reflect who is genuinely available at 3am.
Past Incidents and Issued Statements
What the client has faced before and the language they used publicly. Scenarios get re-scored against what has actually happened to this brand, and new statements stay consistent with positions the client has already taken.
Brand Voice Guide
How the client sounds, and what they never say. A crisis statement still has to sound like the brand, and this keeps the holding lines from reading like boilerplate issued by a different company.
Add a short description to each knowledge item in the project's info field so Juma knows what each file holds and when to reach for it. For example:
A complete playbook covers six parts: a claims exposure audit tied to the brand's own published pages, a scenario register scored on a printed likelihood and impact rubric, a four-level escalation matrix with triggers and update cadence, a first-60-minutes sheet, holding statements written per channel, and a do-not-say table giving the reason behind each forbidden line.
The two deliverables split by use. The branded PDF is the document that gets circulated, reviewed and signed off, with the register, matrix, statements and governance schedule laid out in reading order. The CSV register is the working file: one row per scenario carrying its likelihood and impact scores, risk score, default escalation level, accountable lead, escalation trigger, first actions, statement text and the source each scenario was derived from. Teams sort and filter that file at the quarterly review and re-score rows as the business changes.
A template gives every brand the same scenarios: contamination, cyber incident, executive misconduct. This Flow starts from what the brand has actually published, so the scenarios that matter most name the real claim, quote the wording, link the page and record the date it was checked. The register describes that brand's exposure rather than a category's.
The difference shows up in the highest-scoring rows. Generic registers rank product safety at the top by reflex. A claims-derived register often ranks a sustainability or labelling challenge above it, because the brand has published a specific figure on a specific page that it would have to defend on demand. Each holding statement then points at the claim it exists to answer, rather than leaving a bracket for someone to fill in while the clock runs.
Crisis playbooks usually take an agency two to three weeks: a claims review, a risk workshop, a drafting round, then legal review. This Flow produces the claims audit, scored register, escalation matrix and the full statement set in one working session, which leaves the time for the part that actually needs people.
That part is the judgement. Which risks the client will genuinely accept, who really has authority to speak at 3am, how much candour the brand can carry in its first public line: those are decisions the client's leadership and their counsel make, and no document settles them. Every holding statement arrives marked pre-approved or counsel required precisely so the review goes where it is needed instead of across all seventeen pages. Strategy, taste, and judgment stay human, and human review on every output is the point of the approval column.
Some do and the playbook says which. Each statement carries an approval status: pre-approved means it can go out once the bracketed facts are verified, counsel required means a named approver clears it first. Safety, data, securities and legal-process scenarios land in counsel required by default, because those carry statutory and disclosure consequences.
Treating every statement as equally risky is what makes real playbooks unusable, because teams either route everything to a lawyer who is asleep or route nothing. Splitting the two gives the comms lead something publishable inside the first hour for low-severity scenarios, while the statements that could create liability stop at a named person. The do-not-say table backs this up by naming the exposure behind each forbidden line, so anyone adapting a statement understands why a phrase is off the table rather than deleting a rule they do not recognise.
Yes. The structure holds and the content changes completely. A consumer brand's register fills with product safety, supply, labelling and sustainability claims. A B2B software company's fills with security incidents, availability and uptime commitments, data processing and subprocessor claims, pricing changes and customer concentration.
The claims audit is what makes the difference, because it reads whatever that company has published. For a software business that means the uptime figure on the status page, the certifications on the trust page, the data residency promise in the security overview and the retention language in the privacy policy, each one a commitment somebody can hold them to in public. The escalation matrix changes shape too: enterprise customers often have contractual notification windows, so the matrix carries those clocks alongside the regulatory ones.
Juma finds the earned coverage on the open web, scores every article against a fixed rubric with a verbatim journalist quote as the evidence, and returns a coverage audit plus an editable tracker.
Name the client and their site. Juma returns a claims register where every claim carries the public source that backs it, that source's publication date and the scope the evidence covers, the approved sentence, the phrasing to avoid, and a paste-ready answer block an AI engine can quote and attribute.
400+ marketing teams trust Juma. ChatGPT, Claude, and Perplexity know why. Ask them.
Ready when you are
Start typing a task and Juma will figure out the rest.
